apply.c

   1 /*
   2  * apply.c
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  *
   6  * This applies patches on top of some (arbitrary) version of the SCM.
   7  *
   8  * NOTE! It does all its work in the index file, and only cares about
   9  * the files in the working directory if you tell it to "merge" the
  10  * patch apply.
  11  *
  12  * Even when merging it always takes the source from the index, and
  13  * uses the working tree as a "branch" for a 3-way merge.
  14  */
  15 #include <ctype.h>
  16
  17 #include "cache.h"
  18
  19 // We default to the merge behaviour, since that's what most people would
  20 // expect.
  21 //
  22 //  --check turns on checking that the working tree matches the
  23 //    files that are being modified, but doesn't apply the patch
  24 //  --stat does just a diffstat, and doesn't actually apply
  25 //  --show-files shows the directory changes
  26 //
  27 static int merge_patch = 1;
  28 static int check_index = 0;
  29 static int write_index = 0;
  30 static int diffstat = 0;
  31 static int check = 0;
  32 static int apply = 1;
  33 static int show_files = 0;
  34 static const char apply_usage[] = "git-apply [--stat] [--check] [--show-files] <patch>";
  35
  36 /*
  37  * For "diff-stat" like behaviour, we keep track of the biggest change
  38  * we've seen, and the longest filename. That allows us to do simple
  39  * scaling.
  40  */
  41 static int max_change, max_len;
  42
  43 /*
  44  * Various "current state", notably line numbers and what
  45  * file (and how) we're patching right now.. The "is_xxxx"
  46  * things are flags, where -1 means "don't know yet".
  47  */
  48 static int linenr = 1;
  49
  50 struct fragment {
  51         unsigned long oldpos, oldlines;
  52         unsigned long newpos, newlines;
  53         const char *patch;
  54         int size;
  55         struct fragment *next;
  56 };
  57
  58 struct patch {
  59         char *new_name, *old_name, *def_name;
  60         unsigned int old_mode, new_mode;
  61         int is_rename, is_copy, is_new, is_delete;
  62         int lines_added, lines_deleted;
  63         struct fragment *fragments;
  64         char *result;
  65         unsigned long resultsize;
  66         struct patch *next;
  67 };
  68
  69 #define CHUNKSIZE (8192)
  70 #define SLOP (16)
  71
  72 static void *read_patch_file(int fd, unsigned long *sizep)
  73 {
  74         unsigned long size = 0, alloc = CHUNKSIZE;
  75         void *buffer = xmalloc(alloc);
  76
  77         for (;;) {
  78                 int nr = alloc - size;
  79                 if (nr < 1024) {
  80                         alloc += CHUNKSIZE;
  81                         buffer = xrealloc(buffer, alloc);
  82                         nr = alloc - size;
  83                 }
  84                 nr = read(fd, buffer + size, nr);
  85                 if (!nr)
  86                         break;
  87                 if (nr < 0) {
  88                         if (errno == EAGAIN)
  89                                 continue;
  90                         die("git-apply: read returned %s", strerror(errno));
  91                 }
  92                 size += nr;
  93         }
  94         *sizep = size;
  95
  96         /*
  97          * Make sure that we have some slop in the buffer
  98          * so that we can do speculative "memcmp" etc, and
  99          * see to it that it is NUL-filled.
 100          */
 101         if (alloc < size + SLOP)
 102                 buffer = xrealloc(buffer, size + SLOP);
 103         memset(buffer + size, 0, SLOP);
 104         return buffer;
 105 }
 106
 107 static unsigned long linelen(const char *buffer, unsigned long size)
 108 {
 109         unsigned long len = 0;
 110         while (size--) {
 111                 len++;
 112                 if (*buffer++ == '\n')
 113                         break;
 114         }
 115         return len;
 116 }
 117
 118 static int is_dev_null(const char *str)
 119 {
 120         return !memcmp("/dev/null", str, 9) && isspace(str[9]);
 121 }
 122
 123 #define TERM_SPACE      1
 124 #define TERM_TAB        2
 125
 126 static int name_terminate(const char *name, int namelen, int c, int terminate)
 127 {
 128         if (c == ' ' && !(terminate & TERM_SPACE))
 129                 return 0;
 130         if (c == '\t' && !(terminate & TERM_TAB))
 131                 return 0;
 132
 133         return 1;
 134 }
 135
 136 static char * find_name(const char *line, char *def, int p_value, int terminate)
 137 {
 138         int len;
 139         const char *start = line;
 140         char *name;
 141
 142         for (;;) {
 143                 char c = *line;
 144
 145                 if (isspace(c)) {
 146                         if (c == '\n')
 147                                 break;
 148                         if (name_terminate(start, line-start, c, terminate))
 149                                 break;
 150                 }
 151                 line++;
 152                 if (c == '/' && !--p_value)
 153                         start = line;
 154         }
 155         if (!start)
 156                 return def;
 157         len = line - start;
 158         if (!len)
 159                 return def;
 160
 161         /*
 162          * Generally we prefer the shorter name, especially
 163          * if the other one is just a variation of that with
 164          * something else tacked on to the end (ie "file.orig"
 165          * or "file~").
 166          */
 167         if (def) {
 168                 int deflen = strlen(def);
 169                 if (deflen < len && !strncmp(start, def, deflen))
 170                         return def;
 171         }
 172
 173         name = xmalloc(len + 1);
 174         memcpy(name, start, len);
 175         name[len] = 0;
 176         free(def);
 177         return name;
 178 }
 179
 180 /*
 181  * Get the name etc info from the --/+++ lines of a traditional patch header
 182  *
 183  * NOTE! This hardcodes "-p1" behaviour in filename detection.
 184  *
 185  * FIXME! The end-of-filename heuristics are kind of screwy. For existing
 186  * files, we can happily check the index for a match, but for creating a
 187  * new file we should try to match whatever "patch" does. I have no idea.
 188  */
 189 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
 190 {
 191         int p_value = 1;
 192         char *name;
 193
 194         first += 4;     // skip "--- "
 195         second += 4;    // skip "+++ "
 196         if (is_dev_null(first)) {
 197                 patch->is_new = 1;
 198                 patch->is_delete = 0;
 199                 name = find_name(second, NULL, p_value, TERM_SPACE | TERM_TAB);
 200                 patch->new_name = name;
 201         } else if (is_dev_null(second)) {
 202                 patch->is_new = 0;
 203                 patch->is_delete = 1;
 204                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 205                 patch->old_name = name;
 206         } else {
 207                 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
 208                 name = find_name(second, name, p_value, TERM_SPACE | TERM_TAB);
 209                 patch->old_name = patch->new_name = name;
 210         }
 211         if (!name)
 212                 die("unable to find filename in patch at line %d", linenr);
 213 }
 214
 215 static int gitdiff_hdrend(const char *line, struct patch *patch)
 216 {
 217         return -1;
 218 }
 219
 220 /*
 221  * We're anal about diff header consistency, to make
 222  * sure that we don't end up having strange ambiguous
 223  * patches floating around.
 224  *
 225  * As a result, gitdiff_{old|new}name() will check
 226  * their names against any previous information, just
 227  * to make sure..
 228  */
 229 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
 230 {
 231         int len;
 232         const char *name;
 233
 234         if (!orig_name && !isnull)
 235                 return find_name(line, NULL, 1, 0);
 236
 237         name = "/dev/null";
 238         len = 9;
 239         if (orig_name) {
 240                 name = orig_name;
 241                 len = strlen(name);
 242                 if (isnull)
 243                         die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
 244         }
 245
 246         if (*name == '/')
 247                 goto absolute_path;
 248
 249         for (;;) {
 250                 char c = *line++;
 251                 if (c == '\n')
 252                         break;
 253                 if (c != '/')
 254                         continue;
 255 absolute_path:
 256                 if (memcmp(line, name, len) || line[len] != '\n')
 257                         break;
 258                 return orig_name;
 259         }
 260         die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
 261         return NULL;
 262 }
 263
 264 static int gitdiff_oldname(const char *line, struct patch *patch)
 265 {
 266         patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
 267         return 0;
 268 }
 269
 270 static int gitdiff_newname(const char *line, struct patch *patch)
 271 {
 272         patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
 273         return 0;
 274 }
 275
 276 static int gitdiff_oldmode(const char *line, struct patch *patch)
 277 {
 278         patch->old_mode = strtoul(line, NULL, 8);
 279         return 0;
 280 }
 281
 282 static int gitdiff_newmode(const char *line, struct patch *patch)
 283 {
 284         patch->new_mode = strtoul(line, NULL, 8);
 285         return 0;
 286 }
 287
 288 static int gitdiff_delete(const char *line, struct patch *patch)
 289 {
 290         patch->is_delete = 1;
 291         patch->old_name = patch->def_name;
 292         return gitdiff_oldmode(line, patch);
 293 }
 294
 295 static int gitdiff_newfile(const char *line, struct patch *patch)
 296 {
 297         patch->is_new = 1;
 298         patch->new_name = patch->def_name;
 299         return gitdiff_newmode(line, patch);
 300 }
 301
 302 static int gitdiff_copysrc(const char *line, struct patch *patch)
 303 {
 304         patch->is_copy = 1;
 305         patch->old_name = find_name(line, NULL, 0, 0);
 306         return 0;
 307 }
 308
 309 static int gitdiff_copydst(const char *line, struct patch *patch)
 310 {
 311         patch->is_copy = 1;
 312         patch->new_name = find_name(line, NULL, 0, 0);
 313         return 0;
 314 }
 315
 316 static int gitdiff_renamesrc(const char *line, struct patch *patch)
 317 {
 318         patch->is_rename = 1;
 319         patch->old_name = find_name(line, NULL, 0, 0);
 320         return 0;
 321 }
 322
 323 static int gitdiff_renamedst(const char *line, struct patch *patch)
 324 {
 325         patch->is_rename = 1;
 326         patch->new_name = find_name(line, NULL, 0, 0);
 327         return 0;
 328 }
 329
 330 static int gitdiff_similarity(const char *line, struct patch *patch)
 331 {
 332         return 0;
 333 }
 334
 335 static int gitdiff_dissimilarity(const char *line, struct patch *patch)
 336 {
 337         return 0;
 338 }
 339
 340 /*
 341  * This is normal for a diff that doesn't change anything: we'll fall through
 342  * into the next diff. Tell the parser to break out.
 343  */
 344 static int gitdiff_unrecognized(const char *line, struct patch *patch)
 345 {
 346         return -1;
 347 }
 348
 349 static char *git_header_name(char *line)
 350 {
 351         int len;
 352         char *name, *second;
 353
 354         /*
 355          * Find the first '/'
 356          */
 357         name = line;
 358         for (;;) {
 359                 char c = *name++;
 360                 if (c == '\n')
 361                         return NULL;
 362                 if (c == '/')
 363                         break;
 364         }
 365
 366         /*
 367          * We don't accept absolute paths (/dev/null) as possibly valid
 368          */
 369         if (name == line+1)
 370                 return NULL;
 371
 372         /*
 373          * Accept a name only if it shows up twice, exactly the same
 374          * form.
 375          */
 376         for (len = 0 ; ; len++) {
 377                 char c = name[len];
 378
 379                 switch (c) {
 380                 default:
 381                         continue;
 382                 case '\n':
 383                         break;
 384                 case '\t': case ' ':
 385                         second = name+len;
 386                         for (;;) {
 387                                 char c = *second++;
 388                                 if (c == '\n')
 389                                         return NULL;
 390                                 if (c == '/')
 391                                         break;
 392                         }
 393                         if (second[len] == '\n' && !memcmp(name, second, len)) {
 394                                 char *ret = xmalloc(len + 1);
 395                                 memcpy(ret, name, len);
 396                                 ret[len] = 0;
 397                                 return ret;
 398                         }
 399                 }
 400         }
 401         return NULL;
 402 }
 403
 404 /* Verify that we recognize the lines following a git header */
 405 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
 406 {
 407         unsigned long offset;
 408
 409         /* A git diff has explicit new/delete information, so we don't guess */
 410         patch->is_new = 0;
 411         patch->is_delete = 0;
 412
 413         /*
 414          * Some things may not have the old name in the
 415          * rest of the headers anywhere (pure mode changes,
 416          * or removing or adding empty files), so we get
 417          * the default name from the header.
 418          */
 419         patch->def_name = git_header_name(line + strlen("diff --git "));
 420
 421         line += len;
 422         size -= len;
 423         linenr++;
 424         for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
 425                 static const struct opentry {
 426                         const char *str;
 427                         int (*fn)(const char *, struct patch *);
 428                 } optable[] = {
 429                         { "@@ -", gitdiff_hdrend },
 430                         { "--- ", gitdiff_oldname },
 431                         { "+++ ", gitdiff_newname },
 432                         { "old mode ", gitdiff_oldmode },
 433                         { "new mode ", gitdiff_newmode },
 434                         { "deleted file mode ", gitdiff_delete },
 435                         { "new file mode ", gitdiff_newfile },
 436                         { "copy from ", gitdiff_copysrc },
 437                         { "copy to ", gitdiff_copydst },
 438                         { "rename old ", gitdiff_renamesrc },
 439                         { "rename new ", gitdiff_renamedst },
 440                         { "rename from ", gitdiff_renamesrc },
 441                         { "rename to ", gitdiff_renamedst },
 442                         { "similarity index ", gitdiff_similarity },
 443                         { "dissimilarity index ", gitdiff_dissimilarity },
 444                         { "", gitdiff_unrecognized },
 445                 };
 446                 int i;
 447
 448                 len = linelen(line, size);
 449                 if (!len || line[len-1] != '\n')
 450                         break;
 451                 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
 452                         const struct opentry *p = optable + i;
 453                         int oplen = strlen(p->str);
 454                         if (len < oplen || memcmp(p->str, line, oplen))
 455                                 continue;
 456                         if (p->fn(line + oplen, patch) < 0)
 457                                 return offset;
 458                         break;
 459                 }
 460         }
 461
 462         return offset;
 463 }
 464
 465 static int parse_num(const char *line, unsigned long *p)
 466 {
 467         char *ptr;
 468
 469         if (!isdigit(*line))
 470                 return 0;
 471         *p = strtoul(line, &ptr, 10);
 472         return ptr - line;
 473 }
 474
 475 static int parse_range(const char *line, int len, int offset, const char *expect,
 476                         unsigned long *p1, unsigned long *p2)
 477 {
 478         int digits, ex;
 479
 480         if (offset < 0 || offset >= len)
 481                 return -1;
 482         line += offset;
 483         len -= offset;
 484
 485         digits = parse_num(line, p1);
 486         if (!digits)
 487                 return -1;
 488
 489         offset += digits;
 490         line += digits;
 491         len -= digits;
 492
 493         *p2 = *p1;
 494         if (*line == ',') {
 495                 digits = parse_num(line+1, p2);
 496                 if (!digits)
 497                         return -1;
 498
 499                 offset += digits+1;
 500                 line += digits+1;
 501                 len -= digits+1;
 502         }
 503
 504         ex = strlen(expect);
 505         if (ex > len)
 506                 return -1;
 507         if (memcmp(line, expect, ex))
 508                 return -1;
 509
 510         return offset + ex;
 511 }
 512
 513 /*
 514  * Parse a unified diff fragment header of the
 515  * form "@@ -a,b +c,d @@"
 516  */
 517 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
 518 {
 519         int offset;
 520
 521         if (!len || line[len-1] != '\n')
 522                 return -1;
 523
 524         /* Figure out the number of lines in a fragment */
 525         offset = parse_range(line, len, 4, " +", &fragment->oldpos, &fragment->oldlines);
 526         offset = parse_range(line, len, offset, " @@", &fragment->newpos, &fragment->newlines);
 527
 528         return offset;
 529 }
 530
 531 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
 532 {
 533         unsigned long offset, len;
 534
 535         patch->is_rename = patch->is_copy = 0;
 536         patch->is_new = patch->is_delete = -1;
 537         patch->old_mode = patch->new_mode = 0;
 538         patch->old_name = patch->new_name = NULL;
 539         for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
 540                 unsigned long nextlen;
 541
 542                 len = linelen(line, size);
 543                 if (!len)
 544                         break;
 545
 546                 /* Testing this early allows us to take a few shortcuts.. */
 547                 if (len < 6)
 548                         continue;
 549
 550                 /*
 551                  * Make sure we don't find any unconnected patch fragmants.
 552                  * That's a sign that we didn't find a header, and that a
 553                  * patch has become corrupted/broken up.
 554                  */
 555                 if (!memcmp("@@ -", line, 4)) {
 556                         struct fragment dummy;
 557                         if (parse_fragment_header(line, len, &dummy) < 0)
 558                                 continue;
 559                         error("patch fragment without header at line %d: %.*s", linenr, len-1, line);
 560                 }
 561
 562                 if (size < len + 6)
 563                         break;
 564
 565                 /*
 566                  * Git patch? It might not have a real patch, just a rename
 567                  * or mode change, so we handle that specially
 568                  */
 569                 if (!memcmp("diff --git ", line, 11)) {
 570                         int git_hdr_len = parse_git_header(line, len, size, patch);
 571                         if (git_hdr_len <= len)
 572                                 continue;
 573                         if (!patch->old_name && !patch->new_name) {
 574                                 if (!patch->def_name)
 575                                         die("git diff header lacks filename information (line %d)", linenr);
 576                                 patch->old_name = patch->new_name = patch->def_name;
 577                         }
 578                         *hdrsize = git_hdr_len;
 579                         return offset;
 580                 }
 581
 582                 /** --- followed by +++ ? */
 583                 if (memcmp("--- ", line,  4) || memcmp("+++ ", line + len, 4))
 584                         continue;
 585
 586                 /*
 587                  * We only accept unified patches, so we want it to
 588                  * at least have "@@ -a,b +c,d @@\n", which is 14 chars
 589                  * minimum
 590                  */
 591                 nextlen = linelen(line + len, size - len);
 592                 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
 593                         continue;
 594
 595                 /* Ok, we'll consider it a patch */
 596                 parse_traditional_patch(line, line+len, patch);
 597                 *hdrsize = len + nextlen;
 598                 linenr += 2;
 599                 return offset;
 600         }
 601         return -1;
 602 }
 603
 604 /*
 605  * Parse a unified diff. Note that this really needs
 606  * to parse each fragment separately, since the only
 607  * way to know the difference between a "---" that is
 608  * part of a patch, and a "---" that starts the next
 609  * patch is to look at the line counts..
 610  */
 611 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
 612 {
 613         int added, deleted;
 614         int len = linelen(line, size), offset;
 615         unsigned long oldlines, newlines;
 616
 617         offset = parse_fragment_header(line, len, fragment);
 618         if (offset < 0)
 619                 return -1;
 620         oldlines = fragment->oldlines;
 621         newlines = fragment->newlines;
 622
 623         if (patch->is_new < 0) {
 624                 patch->is_new =  !oldlines;
 625                 if (!oldlines)
 626                         patch->old_name = NULL;
 627         }
 628         if (patch->is_delete < 0) {
 629                 patch->is_delete = !newlines;
 630                 if (!newlines)
 631                         patch->new_name = NULL;
 632         }
 633
 634         if (patch->is_new != !oldlines)
 635                 return error("new file depends on old contents");
 636         if (patch->is_delete != !newlines) {
 637                 if (newlines)
 638                         return error("deleted file still has contents");
 639                 fprintf(stderr, "** warning: file %s becomes empty but is not deleted\n", patch->new_name);
 640         }
 641
 642         /* Parse the thing.. */
 643         line += len;
 644         size -= len;
 645         linenr++;
 646         added = deleted = 0;
 647         for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
 648                 if (!oldlines && !newlines)
 649                         break;
 650                 len = linelen(line, size);
 651                 if (!len || line[len-1] != '\n')
 652                         return -1;
 653                 switch (*line) {
 654                 default:
 655                         return -1;
 656                 case ' ':
 657                         oldlines--;
 658                         newlines--;
 659                         break;
 660                 case '-':
 661                         deleted++;
 662                         oldlines--;
 663                         break;
 664                 case '+':
 665                         added++;
 666                         newlines--;
 667                         break;
 668                 /* We allow "\ No newline at end of file" */
 669                 case '\\':
 670                         if (len < 12 || memcmp(line, "\\ No newline", 12))
 671                                 return -1;
 672                         break;
 673                 }
 674         }
 675         patch->lines_added += added;
 676         patch->lines_deleted += deleted;
 677         return offset;
 678 }
 679
 680 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
 681 {
 682         unsigned long offset = 0;
 683         struct fragment **fragp = &patch->fragments;
 684
 685         while (size > 4 && !memcmp(line, "@@ -", 4)) {
 686                 struct fragment *fragment;
 687                 int len;
 688
 689                 fragment = xmalloc(sizeof(*fragment));
 690                 memset(fragment, 0, sizeof(*fragment));
 691                 len = parse_fragment(line, size, patch, fragment);
 692                 if (len <= 0)
 693                         die("corrupt patch at line %d", linenr);
 694
 695                 fragment->patch = line;
 696                 fragment->size = len;
 697
 698                 *fragp = fragment;
 699                 fragp = &fragment->next;
 700
 701                 offset += len;
 702                 line += len;
 703                 size -= len;
 704         }
 705         return offset;
 706 }
 707
 708 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
 709 {
 710         int hdrsize, patchsize;
 711         int offset = find_header(buffer, size, &hdrsize, patch);
 712
 713         if (offset < 0)
 714                 return offset;
 715
 716         patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
 717
 718         return offset + hdrsize + patchsize;
 719 }
 720
 721 const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
 722 const char minuses[]= "----------------------------------------------------------------------";
 723
 724 static void show_stats(struct patch *patch)
 725 {
 726         char *name = patch->old_name;
 727         int len, max, add, del, total;
 728
 729         if (!name)
 730                 name = patch->new_name;
 731
 732         /*
 733          * "scale" the filename
 734          */
 735         len = strlen(name);
 736         max = max_len;
 737         if (max > 50)
 738                 max = 50;
 739         if (len > max)
 740                 name += len - max;
 741         len = max;
 742
 743         /*
 744          * scale the add/delete
 745          */
 746         max = max_change;
 747         if (max + len > 70)
 748                 max = 70 - len;
 749
 750         add = patch->lines_added;
 751         del = patch->lines_deleted;
 752         total = add + del;
 753
 754         if (max_change > 0) {
 755                 total = (total * max + max_change / 2) / max_change;
 756                 add = (add * max + max_change / 2) / max_change;
 757                 del = total - add;
 758         }
 759         printf(" %-*s |%5d %.*s%.*s\n",
 760                 len, name, patch->lines_added + patch->lines_deleted,
 761                 add, pluses, del, minuses);
 762 }
 763
 764 static int read_old_data(struct stat *st, const char *path, void *buf, unsigned long size)
 765 {
 766         int fd;
 767         unsigned long got;
 768
 769         switch (st->st_mode & S_IFMT) {
 770         case S_IFLNK:
 771                 return readlink(path, buf, size);
 772         case S_IFREG:
 773                 fd = open(path, O_RDONLY);
 774                 if (fd < 0)
 775                         return error("unable to open %s", path);
 776                 got = 0;
 777                 for (;;) {
 778                         int ret = read(fd, buf + got, size - got);
 779                         if (ret < 0) {
 780                                 if (errno == EAGAIN)
 781                                         continue;
 782                                 break;
 783                         }
 784                         if (!ret)
 785                                 break;
 786                         got += ret;
 787                 }
 788                 close(fd);
 789                 return got;
 790
 791         default:
 792                 return -1;
 793         }
 794 }
 795
 796 static int find_offset(const char *buf, unsigned long size, const char *fragment, unsigned long fragsize, int line)
 797 {
 798         int i;
 799         unsigned long start, backwards, forwards;
 800
 801         if (fragsize > size)
 802                 return -1;
 803
 804         start = 0;
 805         if (line > 1) {
 806                 unsigned long offset = 0;
 807                 i = line-1;
 808                 while (offset + fragsize <= size) {
 809                         if (buf[offset++] == '\n') {
 810                                 start = offset;
 811                                 if (!--i)
 812                                         break;
 813                         }
 814                 }
 815         }
 816
 817         /* Exact line number? */
 818         if (!memcmp(buf + start, fragment, fragsize))
 819                 return start;
 820
 821         /*
 822          * There's probably some smart way to do this, but I'll leave
 823          * that to the smart and beautiful people. I'm simple and stupid.
 824          */
 825         backwards = start;
 826         forwards = start;
 827         for (i = 0; ; i++) {
 828                 unsigned long try;
 829                 int n;
 830
 831                 /* "backward" */
 832                 if (i & 1) {
 833                         if (!backwards) {
 834                                 if (forwards + fragsize > size)
 835                                         break;
 836                                 continue;
 837                         }
 838                         do {
 839                                 --backwards;
 840                         } while (backwards && buf[backwards-1] != '\n');
 841                         try = backwards;
 842                 } else {
 843                         while (forwards + fragsize <= size) {
 844                                 if (buf[forwards++] == '\n')
 845                                         break;
 846                         }
 847                         try = forwards;
 848                 }
 849
 850                 if (try + fragsize > size)
 851                         continue;
 852                 if (memcmp(buf + try, fragment, fragsize))
 853                         continue;
 854                 n = (i >> 1)+1;
 855                 if (i & 1)
 856                         n = -n;
 857                 fprintf(stderr, "Fragment applied at offset %d\n", n);
 858                 return try;
 859         }
 860
 861         /*
 862          * We should start searching forward and backward.
 863          */
 864         return -1;
 865 }
 866
 867 struct buffer_desc {
 868         char *buffer;
 869         unsigned long size;
 870         unsigned long alloc;
 871 };
 872
 873 static int apply_one_fragment(struct buffer_desc *desc, struct fragment *frag)
 874 {
 875         char *buf = desc->buffer;
 876         const char *patch = frag->patch;
 877         int offset, size = frag->size;
 878         char *old = xmalloc(size);
 879         char *new = xmalloc(size);
 880         int oldsize = 0, newsize = 0;
 881
 882         while (size > 0) {
 883                 int len = linelen(patch, size);
 884                 int plen;
 885
 886                 if (!len)
 887                         break;
 888
 889                 /*
 890                  * "plen" is how much of the line we should use for
 891                  * the actual patch data. Normally we just remove the
 892                  * first character on the line, but if the line is
 893                  * followed by "\ No newline", then we also remove the
 894                  * last one (which is the newline, of course).
 895                  */
 896                 plen = len-1;
 897                 if (len > size && patch[len] == '\\')
 898                         plen--;
 899                 switch (*patch) {
 900                 case ' ':
 901                 case '-':
 902                         memcpy(old + oldsize, patch + 1, plen);
 903                         oldsize += plen;
 904                         if (*patch == '-')
 905                                 break;
 906                 /* Fall-through for ' ' */
 907                 case '+':
 908                         memcpy(new + newsize, patch + 1, plen);
 909                         newsize += plen;
 910                         break;
 911                 case '@': case '\\':
 912                         /* Ignore it, we already handled it */
 913                         break;
 914                 default:
 915                         return -1;
 916                 }
 917                 patch += len;
 918                 size -= len;
 919         }
 920
 921         offset = find_offset(buf, desc->size, old, oldsize, frag->newpos);
 922         if (offset >= 0) {
 923                 int diff = newsize - oldsize;
 924                 unsigned long size = desc->size + diff;
 925                 unsigned long alloc = desc->alloc;
 926
 927                 if (size > alloc) {
 928                         alloc = size + 8192;
 929                         desc->alloc = alloc;
 930                         buf = xrealloc(buf, alloc);
 931                         desc->buffer = buf;
 932                 }
 933                 desc->size = size;
 934                 memmove(buf + offset + newsize, buf + offset + oldsize, size - offset - newsize);
 935                 memcpy(buf + offset, new, newsize);
 936                 offset = 0;
 937         }
 938
 939         free(old);
 940         free(new);
 941         return offset;
 942 }
 943
 944 static int apply_fragments(struct buffer_desc *desc, struct patch *patch)
 945 {
 946         struct fragment *frag = patch->fragments;
 947
 948         while (frag) {
 949                 if (apply_one_fragment(desc, frag) < 0)
 950                         return error("patch failed: %s:%d", patch->old_name, frag->oldpos);
 951                 frag = frag->next;
 952         }
 953         return 0;
 954 }
 955
 956 static int apply_data(struct patch *patch, struct stat *st)
 957 {
 958         char *buf;
 959         unsigned long size, alloc;
 960         struct buffer_desc desc;
 961
 962         size = 0;
 963         alloc = 0;
 964         buf = NULL;
 965         if (patch->old_name) {
 966                 size = st->st_size;
 967                 alloc = size + 8192;
 968                 buf = xmalloc(alloc);
 969                 if (read_old_data(st, patch->old_name, buf, alloc) != size)
 970                         return error("read of %s failed", patch->old_name);
 971         }
 972
 973         desc.size = size;
 974         desc.alloc = alloc;
 975         desc.buffer = buf;
 976         if (apply_fragments(&desc, patch) < 0)
 977                 return -1;
 978         patch->result = desc.buffer;
 979         patch->resultsize = desc.size;
 980
 981         if (patch->is_delete && patch->resultsize)
 982                 return error("removal patch leaves file contents");
 983
 984         return 0;
 985 }
 986
 987 static int check_patch(struct patch *patch)
 988 {
 989         struct stat st;
 990         const char *old_name = patch->old_name;
 991         const char *new_name = patch->new_name;
 992
 993         if (old_name) {
 994                 int changed;
 995
 996                 if (lstat(old_name, &st) < 0)
 997                         return error("%s: %s", old_name, strerror(errno));
 998                 if (check_index) {
 999                         int pos = cache_name_pos(old_name, strlen(old_name));
1000                         if (pos < 0)
1001                                 return error("%s: does not exist in index", old_name);
1002                         changed = ce_match_stat(active_cache[pos], &st);
1003                         if (changed)
1004                                 return error("%s: does not match index", old_name);
1005                 }
1006                 if (patch->is_new < 0)
1007                         patch->is_new = 0;
1008                 st.st_mode = ntohl(create_ce_mode(st.st_mode));
1009                 if (!patch->old_mode)
1010                         patch->old_mode = st.st_mode;
1011                 if ((st.st_mode ^ patch->old_mode) & S_IFMT)
1012                         return error("%s: wrong type", old_name);
1013                 if (st.st_mode != patch->old_mode)
1014                         fprintf(stderr, "warning: %s has type %o, expected %o\n",
1015                                 old_name, st.st_mode, patch->old_mode);
1016         }
1017
1018         if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
1019                 if (check_index && cache_name_pos(new_name, strlen(new_name)) >= 0)
1020                         return error("%s: already exists in index", new_name);
1021                 if (!lstat(new_name, &st))
1022                         return error("%s: already exists in working directory", new_name);
1023                 if (errno != ENOENT)
1024                         return error("%s: %s", new_name, strerror(errno));
1025                 if (!patch->new_mode)
1026                         patch->new_mode = S_IFREG | 0644;
1027         }
1028
1029         if (new_name && old_name) {
1030                 int same = !strcmp(old_name, new_name);
1031                 if (!patch->new_mode)
1032                         patch->new_mode = patch->old_mode;
1033                 if ((patch->old_mode ^ patch->new_mode) & S_IFMT)
1034                         return error("new mode (%o) of %s does not match old mode (%o)%s%s",
1035                                 patch->new_mode, new_name, patch->old_mode,
1036                                 same ? "" : " of ", same ? "" : old_name);
1037         }
1038
1039         if (apply_data(patch, &st) < 0)
1040                 return error("%s: patch does not apply", old_name);
1041         return 0;
1042 }
1043
1044 static int check_patch_list(struct patch *patch)
1045 {
1046         int error = 0;
1047
1048         for (;patch ; patch = patch->next)
1049                 error |= check_patch(patch);
1050         return error;
1051 }
1052
1053 static void show_file(int c, unsigned int mode, const char *name)
1054 {
1055         printf("%c %o %s\n", c, mode, name);
1056 }
1057
1058 static void show_file_list(struct patch *patch)
1059 {
1060         for (;patch ; patch = patch->next) {
1061                 if (patch->is_rename) {
1062                         show_file('-', patch->old_mode, patch->old_name);
1063                         show_file('+', patch->new_mode, patch->new_name);
1064                         continue;
1065                 }
1066                 if (patch->is_copy || patch->is_new) {
1067                         show_file('+', patch->new_mode, patch->new_name);
1068                         continue;
1069                 }
1070                 if (patch->is_delete) {
1071                         show_file('-', patch->old_mode, patch->old_name);
1072                         continue;
1073                 }
1074                 if (patch->old_mode && patch->new_mode && patch->old_mode != patch->new_mode) {
1075                         printf("M %o:%o %s\n", patch->old_mode, patch->new_mode, patch->old_name);
1076                         continue;
1077                 }
1078                 printf("M %o %s\n", patch->old_mode, patch->old_name);
1079         }
1080 }
1081
1082 static void stat_patch_list(struct patch *patch)
1083 {
1084         int files, adds, dels;
1085
1086         for (files = adds = dels = 0 ; patch ; patch = patch->next) {
1087                 files++;
1088                 adds += patch->lines_added;
1089                 dels += patch->lines_deleted;
1090                 show_stats(patch);
1091         }
1092
1093         printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
1094 }
1095
1096 static void patch_stats(struct patch *patch)
1097 {
1098         int lines = patch->lines_added + patch->lines_deleted;
1099
1100         if (lines > max_change)
1101                 max_change = lines;
1102         if (patch->old_name) {
1103                 int len = strlen(patch->old_name);
1104                 if (len > max_len)
1105                         max_len = len;
1106         }
1107         if (patch->new_name) {
1108                 int len = strlen(patch->new_name);
1109                 if (len > max_len)
1110                         max_len = len;
1111         }
1112 }
1113
1114 static void remove_file(struct patch *patch)
1115 {
1116         if (write_index) {
1117                 if (remove_file_from_cache(patch->old_name) < 0)
1118                         die("unable to remove %s from index", patch->old_name);
1119         }
1120         unlink(patch->old_name);
1121 }
1122
1123 static void add_index_file(const char *path, unsigned mode, void *buf, unsigned long size)
1124 {
1125         struct stat st;
1126         struct cache_entry *ce;
1127         int namelen = strlen(path);
1128         unsigned ce_size = cache_entry_size(namelen);
1129
1130         if (!write_index)
1131                 return;
1132
1133         ce = xmalloc(ce_size);
1134         memset(ce, 0, ce_size);
1135         memcpy(ce->name, path, namelen);
1136         ce->ce_mode = create_ce_mode(mode);
1137         ce->ce_flags = htons(namelen);
1138         if (lstat(path, &st) < 0)
1139                 die("unable to stat newly created file %s", path);
1140         fill_stat_cache_info(ce, &st);
1141         if (write_sha1_file(buf, size, "blob", ce->sha1) < 0)
1142                 die("unable to create backing store for newly created file %s", path);
1143         if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD) < 0)
1144                 die("unable to add cache entry for %s", path);
1145 }
1146
1147 static void create_subdirectories(const char *path)
1148 {
1149         int len = strlen(path);
1150         char *buf = xmalloc(len + 1);
1151         const char *slash = path;
1152
1153         while ((slash = strchr(slash+1, '/')) != NULL) {
1154                 len = slash - path;
1155                 memcpy(buf, path, len);
1156                 buf[len] = 0;
1157                 if (mkdir(buf, 0755) < 0) {
1158                         if (errno != EEXIST)
1159                                 break;
1160                 }
1161         }
1162         free(buf);
1163 }
1164
1165 /*
1166  * We optimistically assume that the directories exist,
1167  * which is true 99% of the time anyway. If they don't,
1168  * we create them and try again.
1169  */
1170 static int create_regular_file(const char *path, unsigned int mode)
1171 {
1172         int ret = open(path, O_WRONLY | O_CREAT | O_TRUNC, mode);
1173
1174         if (ret < 0 && errno == ENOENT) {
1175                 create_subdirectories(path);
1176                 ret = open(path, O_WRONLY | O_CREAT | O_TRUNC, mode);
1177         }
1178         return ret;
1179 }
1180
1181 static int create_symlink(const char *buf, const char *path)
1182 {
1183         int ret = symlink(buf, path);
1184
1185         if (ret < 0 && errno == ENOENT) {
1186                 create_subdirectories(path);
1187                 ret = symlink(buf, path);
1188         }
1189         return ret;
1190 }
1191
1192 static void create_file(struct patch *patch)
1193 {
1194         const char *path = patch->new_name;
1195         unsigned mode = patch->new_mode;
1196         unsigned long size = patch->resultsize;
1197         char *buf = patch->result;
1198
1199         if (!mode)
1200                 mode = S_IFREG | 0644;
1201         if (S_ISREG(mode)) {
1202                 int fd;
1203                 mode = (mode & 0100) ? 0777 : 0666;
1204                 fd = create_regular_file(path, mode);
1205                 if (fd < 0)
1206                         die("unable to create file %s (%s)", path, strerror(errno));
1207                 if (write(fd, buf, size) != size)
1208                         die("unable to write file %s", path);
1209                 close(fd);
1210                 add_index_file(path, mode, buf, size);
1211                 return;
1212         }
1213         if (S_ISLNK(mode)) {
1214                 if (size && buf[size-1] == '\n')
1215                         size--;
1216                 buf[size] = 0;
1217                 if (create_symlink(buf, path) < 0)
1218                         die("unable to write symlink %s", path);
1219                 add_index_file(path, mode, buf, size);
1220                 return;
1221         }
1222         die("unable to write file mode %o", mode);
1223 }
1224
1225 static void write_out_one_result(struct patch *patch)
1226 {
1227         if (patch->is_delete > 0) {
1228                 remove_file(patch);
1229                 return;
1230         }
1231         if (patch->is_new > 0 || patch->is_copy) {
1232                 create_file(patch);
1233                 return;
1234         }
1235         /*
1236          * Rename or modification boils down to the same
1237          * thing: remove the old, write the new
1238          */
1239         remove_file(patch);
1240         create_file(patch);
1241 }
1242
1243 static void write_out_results(struct patch *list)
1244 {
1245         if (!list)
1246                 die("No changes");
1247
1248         while (list) {
1249                 write_out_one_result(list);
1250                 list = list->next;
1251         }
1252 }
1253
1254 static struct cache_file cache_file;
1255
1256 static int apply_patch(int fd)
1257 {
1258         int newfd;
1259         unsigned long offset, size;
1260         char *buffer = read_patch_file(fd, &size);
1261         struct patch *list = NULL, **listp = &list;
1262
1263         if (!buffer)
1264                 return -1;
1265         offset = 0;
1266         while (size > 0) {
1267                 struct patch *patch;
1268                 int nr;
1269
1270                 patch = xmalloc(sizeof(*patch));
1271                 memset(patch, 0, sizeof(*patch));
1272                 nr = parse_chunk(buffer + offset, size, patch);
1273                 if (nr < 0)
1274                         break;
1275                 patch_stats(patch);
1276                 *listp = patch;
1277                 listp = &patch->next;
1278                 offset += nr;
1279                 size -= nr;
1280         }
1281
1282         newfd = -1;
1283         write_index = check_index && apply;
1284         if (write_index)
1285                 newfd = hold_index_file_for_update(&cache_file, get_index_file());
1286         if (check_index) {
1287                 if (read_cache() < 0)
1288                         die("unable to read index file");
1289         }
1290
1291         if ((check || apply) && check_patch_list(list) < 0)
1292                 exit(1);
1293
1294         if (apply)
1295                 write_out_results(list);
1296
1297         if (write_index) {
1298                 if (write_cache(newfd, active_cache, active_nr) ||
1299                     commit_index_file(&cache_file))
1300                         die("Unable to write new cachefile");
1301         }
1302
1303         if (show_files)
1304                 show_file_list(list);
1305
1306         if (diffstat)
1307                 stat_patch_list(list);
1308
1309         free(buffer);
1310         return 0;
1311 }
1312
1313 int main(int argc, char **argv)
1314 {
1315         int i;
1316         int read_stdin = 1;
1317
1318         for (i = 1; i < argc; i++) {
1319                 const char *arg = argv[i];
1320                 int fd;
1321
1322                 if (!strcmp(arg, "-")) {
1323                         apply_patch(0);
1324                         read_stdin = 0;
1325                         continue;
1326                 }
1327                 if (!strcmp(arg, "--no-merge")) {
1328                         merge_patch = 0;
1329                         continue;
1330                 }
1331                 if (!strcmp(arg, "--stat")) {
1332                         apply = 0;
1333                         diffstat = 1;
1334                         continue;
1335                 }
1336                 if (!strcmp(arg, "--check")) {
1337                         apply = 0;
1338                         check = 1;
1339                         continue;
1340                 }
1341                 if (!strcmp(arg, "--index")) {
1342                         check_index = 1;
1343                         continue;
1344                 }
1345                 if (!strcmp(arg, "--show-files")) {
1346                         show_files = 1;
1347                         continue;
1348                 }
1349                 fd = open(arg, O_RDONLY);
1350                 if (fd < 0)
1351                         usage(apply_usage);
1352                 read_stdin = 0;
1353                 apply_patch(fd);
1354                 close(fd);
1355         }
1356         if (read_stdin)
1357                 apply_patch(0);
1358         return 0;
1359 }